Detecting slow slip signals in southwest Japan based on machine learning trained by real GNSS time series
- Yusuke Tanaka
, - Masayuki Kano,
- Keisuke Yano
Masayuki Kano
Graduate School of Science, Solid Earth Physics Laboratory, Tohoku University
Keisuke Yano
The Institute of Statistical Mathematics
Abstract
• We succeeded single-site signal detection based on real displacement time series of shortterm slow slip events in southwest Japan. • We obtained detection accuracy of 97-98% with consistent spatiotemporal trend of performance compared to the existing detection catalog. • We could directly capture the complex spatio-temporal variation of real noise and discussed its influence on the detection performance.24 May 2024Submitted to ESS Open Archive 28 May 2024Published in ESS Open Archive
